HCDC vs NIHE data

Compare births from HCDC vs Registry

Some findings:

• NIHE data have higher birth counts in 2016 - 2018 which contributes to lower coverage (2022, 2014 low counts are due to incomplete data collection). This maybe an indication that the threshold for duplicate matching is too high (but lowering it may impact result for 2019-2021)

• Discrepancies diminishes overtime (NIHE even slightly under under reported number of births for some districts by 2021)

• Over-reported (compared to HCDC) birth counts in NIHE happen accross all districts (except Hoc Mon)

All

By district

Click on the district to view number of births over the year for that district

Note: \(Ratio = \frac{\text{Total births from NIHE}}{\text{Total births from HCDC data}}\)

`summarise()` has grouped output by 'namsinh'. You can override using the
`.groups` argument.
`summarise()` has grouped output by 'birth_year'. You can override using the
`.groups` argument.

Data table

rmarkdown::paged_table(count_by_district %>% select(-joinkey))

Compare vaccinated children count from HCDC vs Registry

Some findings:

• Records from NIHE show lower vaccinated population compared to HCDC data, especially in 2016-2018 and 2021 which further contribute to the discrepancies in coverage for those years in particular
• Similar to birth counts, the discrepancies between HCDC and NIHE count can be found across all districts
• The vaccinated population for District 1 is the most consistent between HCDC and NIHE

All

By district

Click on the district to view number of vaccinated children over the year for that district

Note:

\(Ratio = \frac{\text{Vaccinated children from NIHE}}{\text{Vaccinated children from HCDC data}}\)

`summarise()` has grouped output by 'birth_year'. You can override using the
`.groups` argument.

Data table

rmarkdown::paged_table(vacc_by_district %>% select(-joinkey))

NIHE coverage vs HCDC coverage

Discrepancy by District

Discrepancy between coverage calculated from HCDC and NIHE is consistent across all districts

`summarise()` has grouped output by 'district'. You can override using the
`.groups` argument.

Data table

Geo plot for coverage computed from HCDC data

Map for Thu Duc

I had some bug when trying to load map with merged Thu Duc city which have not been resolved so the coverage for Thu Duc in the following map is in fact coverage for Thu Duc district, District 2 and District 9

import {addTooltips} from "@mkfreeman/plot-tooltip"
Plot = import("https://cdn.jsdelivr.net/npm/@observablehq/plot/+esm")
percent = d3.format(".2%")

district_map = FileAttachment("data/gadm41_VNM_2.json").json();
// only keep district of Ho Chi Minh city
district_map.features = district_map.features.filter(
 d => d.properties.NAME_1 == "HồChíMinh"
)

hcdc_coverage = FileAttachment("data/hcdc_coverage_data.csv").csv({typed: true});

viewof year_select = Inputs.range([2016, 2024], 
                                 {label: "Year of birth:", step: 1, value: 2024})

filtered_cov = hcdc_coverage.filter((d) => d.namsinh === year_select)
// create map from district to coverage
district_cov_map = new Map(
  filtered_cov.map( 
    // fix names to mix that of map data
  (d) => [
    d.qh.normalize('NFD').replace(/[\u02c6-\u0323]/g, '')
    .replace("Đ", "D").replace("Quan", "District").toLowerCase().
    split(' ').join(''), 
  d.coverage] ))

addTooltips(
  Plot.plot({
    projection: {type: "mercator", domain: district_map},
    color: {
        scheme: "blues",
        label: "Coverage (%)",
        legend: true,
        percent: true, // Convert value to a percent (from a proportion)
        domain: [0, 100] // Update the value domain to span 0 to 100% access
      },
    marks: [
      Plot.geo(district_map, 
      {
        fill: (d) => district_cov_map.get(
          // update district name to match that in the registry for mapping
          d.properties.VARNAME_2.toLowerCase()
          ),
        title: (d) => `${d.properties.NAME_2}
                    ${percent(district_cov_map.get(d.properties.VARNAME_2.toLowerCase()))}`
      }),
      Plot.geo(district_map, {stroke: "black", strokeWidth: 0.5})
    ]
  })
)